Here’s Llama 3.3 70B checking events in San Francisco.
— Aaron Ng (@localghost) 20 février 2025
Every AI can use the internet now. pic.twitter.com/eJ804KxJuO
Here’s Llama 3.3 70B checking events in San Francisco. Every AI can use the internet now.
@localghost
-
Llama 3.3 70B Now Has Internet Access for Event Checking
By
–
-
OpenAI Strategy: Model Releases and Open Source o3-mini Alternative
By
–
they don't usually release models unless they've got a strong lead in some dimension. but would rather see oss o3-mini if they can't do that.
-
Phone AI Model Performance Threshold for Product Innovation
By
–
A 10x better phone AI model completely changes what people can build. If it's just 20% better it won't matter at all and they should just oss o3-mini instead. So the specifics matter here.
-
Tiny AI Models Need Quantum Leaps Not Incremental Gains
By
–
the base itself needs to be a leap ahead though or there's no point. a 2-10X better tiny model would be amazing. a 10% better tiny model would be pointless
-
Building the First Tiny Smart Model Not Another Generic One
By
–
no point in doing it if it's the 101st tiny dumb model, it needs to be the 1st tiny smart model
-
Small Base Models Beyond Distillation: Seeking Better Alternatives
By
–
id want to see a great small base model. i don't think distilling alone will yield something a few multiples better than what's out there today
-
O3-Mini distillation challenges depend on student model architecture
By
–
a distillation being good will still depend on the student model architecture. not convinced distilling o3-mini to existing models will yield anything significantly better than what's out there today
-
Open-source o3-mini models and small local AI advancement
By
–
Phone model. Here's why: 1. An o3-mini-level oss model is short-term inevitable. It will happen from somewhere, and it will happen soon. 2. Small local models still have a long way to go. A truly great <8B model will unlock many new uses: edge devices, small projects, etc.
-
Microsoft Accelerates 250 Years Chemistry Progress Into 25
By
–
"Our goal is to compress the next 250 years of chemistry and materials science progress into the next 25." Satya said this in 2023. Looks like they might actually do it.
-
Apollo Adds Re-roll Feature for Generated Replies
By
–
You can now re-roll replies in Apollo.
— Aaron Ng (@localghost) 19 février 2025
Long-time request, out in the latest version. pic.twitter.com/N8AdrXEkHjYou can now re-roll replies in Apollo. Long-time request, out in the latest version.